Model: gemini-3.1-flash-live-preview (and the GA gemini-3.1-flash-live id when available)
GCP Project ID: *****-v0ice-prod
Region: global (and us-central1)
Billing: enabled and active
Vertex AI API: aiplatform.googleapis.com enabled
Org policy: constraints/vertexai.allowedModels effective policy is allValues: ALLOW
Use case: Production inbound/outbound voice agent, live speech-to-speech over the Live API (bidirectional WebSocket, PCM 16k in / 24k out).
Why now: We currently run gemini-live-2.5-flash-native-audio, retiring 2026-12-13. Google’s retirement notice names gemini-3.1-flash-live as the successor. models.list() on this project returns only three Live models — gemini-live-2.5-flash-native-audio, gemini-3.5-live-translate-preview, gemini-3.5-transcribe-live-preview — the latter two being translation/transcription endpoints, not conversational. So there is currently no successor available to migrate to.